Research map / Probabilistic and scientific modeling / Molecules and proteins
Multi-marginal temporal Schrödinger Bridge Matching from unpaired data
Thomas Gravier, Thomas Boyer, Auguste Genovesio
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use MMtSBM instead of the previous best comparable SB method [SF]2M-Geo | MMtSBM · MULTI d=100 | W1: 52.203 -> 44.542 | Improved |
| use MMtSBM instead of DMSB | MMtSBM · Embryoid Body d=100 (all marginals) | MMD: 0.032 -> 0.019 | Improved |
| use MMtSBM instead of DMSB | MMtSBM · Embryoid Body d=100 (all marginals) | SWD: 0.16 -> 0.130 | Improved |
| use MMtSBM instead of 3MSBM on held-out times | MMtSBM · Embryoid Body d=100 (held-out) | MMD t3: 0.14 -> 0.06 | Improved |
| add SB learning phase (IMFF) after flow-matching warm-up (ablation) | MMtSBM · 2D Gaussian mixtures N=3 | trajectory optimality | Improved |
| use naive (non-factorized) multi-marginal IMF (ablation) | IMF extension · multi-marginal convergence | convergence | Hurt |
| train multi-marginal bridge instead of single bridge | MMtSBM · Moons (N=4) | W2: 0.144 -> 0.148 | No clear effect |
| run SB phase after warm-up in 50d Gaussian transport (ablation) | MMtSBM · 50d Gaussian multi-marginal | covariance | Improved |
Closest papers by meaning
- Momentum Multi-Marginal Schrödinger Bridge Matching (NeurIPS 2025 · similarity 0.93)
- Trajectory Inference with Smooth Schrödinger Bridges (ICML 2025 · similarity 0.89)
- Learning non-equilibrium diffusions with Schrödinger bridges: from exactly solvable to simulation-free (NeurIPS 2025 · similarity 0.89)
- Feedback Schrödinger Bridge Matching (ICLR 2025 · similarity 0.88)
- Generalized Schrödinger Bridge Matching (ICLR 2024 · similarity 0.88)
- Modeling Cell Dynamics and Interactions with Unbalanced Mean Field Schrödinger Bridge (NeurIPS 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
